Products such as semiconductor or liquid crystal displays are manufactured by a series of process steps including cleaning, resist coating, exposure, development, etching, formation of an interlayer insulating film, thermal processing on substrates. These process steps are generally performed in a substrate processing apparatus incorporating a plurality of processing units. In the substrate processing apparatus, a transport robot sequentially transports a substrate between the processing units at which a substrate is subjected to a predetermined process to experience a series of process steps.
Conventionally, these processing units, transport robot and the like are each provided with various types of sensors for monitoring operations and process conditions thereof. These sensors detect an abnormality in substrate processing, an exemplary technique of which is disclosed in Japanese Patent Application Laid-Open No. 11-3117. Various types of sensors monitor respective targets. When any one of the respective targets goes out of a predetermined range, a corresponding one of the sensors sends an abnormality detecting signal so that a processing abnormality is detected.
An apparatus for removing polymer is an exemplary substrate processing apparatus in which polymers deposited on a substrate which are generated as a result of transformation of a resist in dry etching using reactive ions are removed by cleaning. In such a polymer removal apparatus, sensors detect the number of revolutions of a spin motor for rotating a substrate, the flow rate of a polymer removal solution and the like. In the event that any one of these values goes out of a predetermined range, a corresponding sensor sends an abnormality detecting signal.
However, a combination of a plurality of factors synthetically acts to obtain the results of substrate processing. For example, the number of revolutions of a spin motor even going out of a predetermined range may cause no abnormality of the results of processing as a result of combination with other control elements.
Conversely, even when all control elements fall within their predetermined ranges, a plurality of control elements as a combination may cause an abnormality of the results of processing. In this case, no abnormality detecting signal is sent from each sensor and hence, a processing abnormality should be detected in an inspection step to be performed after a series of substrate process steps is completed.
That is, even with various types of sensors provided to each processing unit, transport robot and the like, processing giving rise to no problem as a whole may be detected as a processing abnormality, or processing containing an abnormality may not be detected as an abnormality. Thus, it has been difficult to detect a processing abnormality with a high degree of accuracy by monitoring the operation of each processing unit.